Download Appendix S2 File

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Bootstrapping (statistics) wikipedia , lookup

Taylor's law wikipedia , lookup

Psychometrics wikipedia , lookup

Student's t-test wikipedia , lookup

Regression toward the mean wikipedia , lookup

Resampling (statistics) wikipedia , lookup

Transcript
Appendix S2 File: Environmental/topographic heterogeneity of
both continents
1. MATERIALS AND METHODS
We calculated summary statistics with the topographic information on the temperate regions in
Europe and North America. We chose two approaches to quantify environmental heterogeneity:
1) Since the analyzed regions on both continents cover different extents of landmass, we sampled
both continents to the same number of grid cells 999 times and used Wilcoxon signed-rank tests
to test for differences between the sampled mean values and standard deviations in altitude
between the continents. We used bootstrap tests with 999 permutations on the W-statistics to
infer the central tendencies and the robustness from these samples. 2) Altitudinal maps have a
resolution of 5 arc min which equals an area of app. 30 x 30 km close to the equator. We
aggregated these high resolution maps to the resolution of 1 degree which is app. 100 x 100 km
close to the equator (3 x 3 grid cells of 5 arc min resolution form one grid cell of 1 degree
resolution) and used the standard deviation of the aggregated grid cells as a measure of spatial
heterogeneity. We sampled the standard deviation-values of both continents to the same number
and compared the mean values with Wilcoxon signed-rank tests. We used bootstrap tests with
999 permutations on the W-statistics to infer the central tendencies and the robustness of these
samples.
2. RESULTS
summary statistics
The sample of the North American temperate region is on average 20 m (mean) to 80 m (median)
higher in altitude than the European sample (
Table 1). The distribution of altitudes differs between both continents (Figure 1).
Table 1 Summary statistics of topographic information on the analyzed regions.
Min [m]
1st Quantile [m]
Mean [m]
Median [m]
3rd Quantile [m]
Europe
-6
70
198
142
254
North America
-22
121
223.5
218
311
Figure 1 Histograms and Box-Whisker-Plots for the analyzed regions.
First test for environmental heterogeneity
The mean altitude of grid cells (5 arc min resolution) is significantly higher in Europe than in North
America in subregion 1 and 2 (Table 1). In subregion 3 and 4, as well as for the whole temperate
region, the mean altitude is higher in North America than in Europe. However, the standard
deviation around these mean values is significantly higher in Europe than in North America for all
subregions and for the whole temperate region (Error! Reference source not found.).
Table 1 First test of environmental heterogeneity for the mean values of altitude in 5 arc min grid cells.
area1
area2
area3
area4
all
regions
mean altitude
Europe [m]
308.9
261.8
112.9
220.1
mean altitude
North America [m]
162.8
248.4
231.2
220.3
198
223.5
permutation test of 999 Wilcoxon-signed rank
tests of differences between the altitude of
both continents
mean Wstatistic
bias
mean St.-Error
4025777
14.85973
995.2463
46759225
-42.00207
2592.142
3151836
0.2489545
55.35026
8259805
10.32182
477.4031
631400000
1053.302
45738.72
Confidence interval of the normal
distribution for the mean W-statistic
2.50%
3828019
46255393
3141226
8165781
97.50%
4223177
47266726
3162680
8354428
significance
p < 0.001
p < 0.001
p < 0.001
p < 0.001
630809085
636310074
p < 0.001
Table 2 First test of environmental heterogeneity for the standard deviation of altitudes in 5 arc min grid cells.
area1
area2
area3
area4
all
regions
sd Europe [m]
124.2447
228.821
119.6064
182.2205
sd North America
[m]
71.37785
107.8128
100.8107
205.6607
191.7573
154.8825
permutation test of 999 Wilcoxon-signed rank
tests of differences between the altitude of
both continents
mean Wstatistic
bias
mean St.-Error
4026422
81.51919
3173.818
46763323
93.06622
7869.331
3152146
-1.220509
169.9644
8261476
56.68444
1526.031
633466939
-262.9277
46831.29
Confidence interval of the normal
distribution for the mean W-statistic
2.50%
3834588
46277710
3141969
8166969
97.50%
4218197
47242226
3162889
8359512
significance
p < 0.001
p < 0.001
p < 0.001
p < 0.001
630645123
636413468
p < 0.001
Second test for environmental heterogeneity
In Europe grid cells of an app. area of 120 x 120 km have an intrinsic altitudinal difference of 66 m. In North America this difference/ heterogeneity is lower
with a mean value of 45 m (Table 2 Second test of environmental heterogeneity for the intrinsic standard deviation of altitudes in grid cells of 1 degree
resolution.).
Table 2 Second test of environmental heterogeneity for the intrinsic standard deviation of altitudes in grid cells of 1 degree resolution.
Mean sd
Europe [m]
Mean sd North
America [m]
66.1
45.0
Temperate
region
permutation test of 999 Wilcoxon-signed
rank tests of differences between the
altitude of both continents
mean Wstatistic
bias
mean St.-Error
51900.84
2.125034
35.17492
Confidence interval of the normal
distribution for the mean W-statistic
2.50%
97.50%
significance
49659.85
54085.2
p < 0.001
4. DISCUSSION
Our results suggest a stronger environmental heterogeneity, as measured by altitude, in Europe than in North America.